2014 kicked off an era of innovation and promise for digital health and Artificial Intelligence (AI) in healthcare. Apple launched HealthKit, Watch, HomeKit, ResearchKit (2015), CareKit (2016); the FDA announced the first mobile and social guidelines; Google acquired DeepMind (2014); and Microsoft continued to evangelize their HealthVault product. The most notable push in AI for in the space, came by the way of IBM and the aggressive push of their AI platform, Watson Health.
Despite IBM Watson Health’s strong trajectory including staffing, conferences, extensive marketing, and key partnerships with such notable organizations as The Mayo Clinic, Epic Systems, and Memorial Sloan Kettering Cancer Center, their efforts have taken somewhat of a downturn. Trending opinions from former partners, industry leaders, and the like suggest that the fundamental gaps and actual delivery of real-world outcomes have outweighed the optimism.
Two Steps Forward, Three Steps Back
While all of this might sound a bit discouraging, like many new technologies and innovations, it often takes these kinds of setbacks to identify the underlying, often unknown components that might not have been revealed otherwise. (After all, desktop computers had no use until the first spreadsheet software was invented.)
The light in the tunnel today is promising in a lot of ways as a result of these insights. However, the light at the end of the tunnel is still some ways away.
One Step Forward, Another Step Forward
Despite all of the commentary on the shortcomings of IBM Watson Health, it can definitely be credited for bringing exposure, even notoriety to the advent and advancement of AI. As such, since 2014 a slew of new entrants, startups, and within known entities (e.g., Microsoft, Google), have sprouted. For example, companies such as Saama creating new mechanisms using AI to scrub and analyze data for clinical trials, patient engagement, and research, or Tellic, which uses AI to help pharma companies accelerate drug development. This is not to suggest the water’s fine and let’s all dive in.
Quite the contrary, innovators across pharma, healthcare professionals, healthcare institutions, data scientists, technologists, medical regulatory, even marketers must tread carefully while taking an even deeper dive into the learnings about AI. These groups need to isolate the critical data and other such touchpoints in framing AI process and infrastructure to establish a fundamental basis for success.
A Brave (but Prudent) New World
As evidenced by the rise of the aforementioned new entrants, the landscape is also experiencing focused conferences and bodies of information about AI in healthcare and life sciences. A recent event from ExL Pharma, “AI Innovations for Life Science and Health Care Summit – June 2018,” featured a diverse audience engaged by speakers from across very differing disciplines and roles, as well as the introduction of new senior roles seemingly born of hybrid skillsets and experiences. Examples of these roles include Summit speakers: Ryan Billings, Senior Director, Digital Innovation at AMAG Pharmaceuticals; Christine Sawicki, Director Specialty Innovation CVS Health; and Akbar Siddiqui, Lead Architect – Cognitive Intelligence at Bristol-Myers Squibb.
Beyond the growth of these roles and general interest in AI, this summit, its attendees, and some other recent news helped to shine a light about where AI is headed and how it is being adapted within pharma.
The insights and information driven by the above speakers at the summit was unique in that they used their hybrid experience to speak on such topics as: “The Implications of the Data Deluge in Pharma,” “Partnerships with Health Systems and Tech to Deliver Better Care” as well as “Best Practices in Decision Management to Ensure the Subject Matter Expert is in Control” to name a notable few.
Another standout element of the AI Summit were the attendees themselves. Senior roles from organizations such as GSK, University of Notre Dame, Pfizer, IEEE Standards Association, and Humana, as well as others were a truly engaged and participatory audience. It was clear from overall attendee commentary that they were there to learn as well as provide insights from their own organizations and experience.
Or, more to the point, breaking it down seems to be the mantra across these organizations and disciplines. The new roles we saw within pharma companies at the summit and a more concerted focus on the newly discovered gaps is revealing even broader learnings and new opportunities. Taking not only a deeper dive into the data, but a dive into how unstructured data typical of natural language fits into the analysis and insights.
Real-World Artificial Intelligence
Deciphering the data as such, is a key restarting point since AI is driven by data. Here, as in marketing, content is also king. Recognizing the decoupling of data typical of regulated environments that can result in still very manual models of data gathering, disparate provider data across a slew of medical health record systems, the need to create methods of connectivity across environments will be predominant tasks necessary towards producing actionable insights.
In pharma, key emerging trends include clinical trial mechanisms, image analysis, digital assistants, even digital therapies. Pfizer in fact has moved from five to 100 AI projects in the last year alone. Before you gasp, bear in mind such projects across organizations have included a longer vetting process with added touchpoints before even consider going to market. A good amount of these projects are therefore, finding an end point in “pilot products.”
Artificial Intelligence by any Other Name is Actually Less Scary
Brands and marketers are shying away from the label AI and moving towards more benign titles such as IA—Intelligence Assistant. This is an important point to drive home: The idea that AI is here to augment and enhance and not replace.
Well, it all needs to be patient centered. Let’s face it, historically the various touchpoints a patient moves through in the medical system are typically disconnected even divergent. The blockchain approach proposes, at least conceptually, to move a patient with greater ease, security, and assurance across their journey of information and interaction in a more connected and secure approach. Even the inclusion of digital assistants will help to drive real human interactions in a space where HCP to patient relationships could be stronger.
The FDA Continues to Step Up
The FDA in April of 2018, released a paper to include AI and digital health in general. “AI holds enormous promise for the future of medicine,” said Scott Gottlieb, MD, Commissioner of Food and Drugs, (2018) “and we’re actively developing a new regulatory framework to promote innovation in this space and support the use of AI-based technologies.” This is a significant opportunity for HCPs, brands, and marketers to follow—even shape—it’s evolution. It will also help to inform how client regulatory might leverage this information as well as how marketers create content for brands.
We asked Scott Hansen, EVP, Creative Director at Stori Health, his thoughts on what AI in this respect would represent in creating brand content.
“When writing for brands, I generally work with published studies. Starting with pivotal trial data. Perhaps augmenting with survey type studies that provide disease-specific content. Finally, looking at market research to try to truly understand needs and niches.”
And as far as AI, “AI provides a platform for Socratic learning. Rather than dictation, AI allows the user to ask questions along their stream of consciousness crafting their own narrative. The user is able to discover what they want to know. This is opposite from me telling you what I want you to know.”
Conclusions are Inconclusive
Despite all of the firsts and starts, this new stage in Artificial Intelligence for healthcare is finding advocacy, effort, and support across organizations and interest groups. It’s a digital-first world, but prudence dictates that it doesn’t need to be a digital-fast world. There are still hurdles and challenges that need to be overcome.
A careful but concerted movement by healthcare into AI will help ensure its usability for brands, marketers, HCPs, and especially patients towards more human insights and potentially better treatment outcomes. Remember the tortoise beat the hare with its slow but steady approach.